Experimental and numerical investigations of the influence of cracks on mass diffusion in mortar and concrete

Even though it was intensely studied in the last decades, chloride ingress into cracked concrete and its influence on reinforced concrete (RC) structures with cracks is still not sufficiently understood, due to its complexity. Along the years, different testing methods were adopted by researchers in order to study the influence of different factors: crack characteristics, environmental parameters, concrete properties and mitigating mechanisms on chloride penetration in cracked concrete. Due to their many advantages, a great interest has been shown in developing a numerical model that can accurately predict chloride penetration into concrete by taking into consideration several aspects of the transport mechanisms of chloride. Still, very limited investigation of the influence of chloride diffusion on samples with real cracks can be found in literature. In this thesis an attempt was made to develop a numerical model that can accurately simulate chloride penetration in structures with real cracks. Chapter 1 presents the introduction, underlining the motivation and the background of the present research project. General information about chloride penetration in concrete is presented such as the service life of a RC structure, the critical chloride value and corrosion mechanisms are presented. Also an outline of the thesis is presented. In chapter 2 a brief review of both the experimental and numerical information found in literature regarding chloride ingress in uncracked and cracked concrete is provided. First, several crack preparation methods are presented, followed by the description of the most common experimental methods used to determine chloride penetration in concrete: migration tests and diffusion tests. The main chloride transport mechanisms are then given and briefly discussed. Also the influence of different parameters (crack width, crack depth, water-to-binder ratio, cement content and cement type, and loading conditions) and mitigating mechanisms on chloride penetration in cracked concrete are presented. Finally, a description of the various numerical and modelling techniques found in literature analysing the two or three- dimensional aspects of chloride ingress in cracked concrete taking into account its different transport mechanisms is provided. Chapter 3 deals with the experimental part of the research. First, an overview of the experimental part is provided. In order to study the influence of real cracks on the chloride penetration in RC structures, 40 cores with 100 mm diameter were drilled from different locations from a previously loaded RC slab, manufactured with a concrete class C30/37 with a maximum size of aggregate of 14 mm. The samples where then prepared to meet the requirements specified in the standard NT BUILD 492 (1999) for the non-steady state migration test (100 mm diameter and 50 mm thickness). Based on their characteristics, the 40 samples were then grouped in four main categories: samples without cracks and without rebars (S), samples with cracks and without rebars (SC), samples without cracks and with rebars (SR), samples with cracks and with rebars (SCR). For each particular sample, geometric characteristics (crack width, crack tortuosity and roughness, rebar position) have been determined, and the position of the carbonation front was also detected. In addition to crack width measurements along the top and bottom surfaces, each core was vertically sectioned in different locations, perpendicular to the surface crack. The crack profile through the specimen's thickness was studied and the crack width through the sample was determined. In order to experimentally determine the chloride ingress, a non-steady state migration test was performed (NT BUILD 492, 1999) and both the chloride penetration profile and the migration coefficient were determined for each sample. Based on the experimental data it was found that the presence of cracks has a significant influence on chloride ingress, increasing it, but also that chloride penetration is significantly affected by the characteristics of the concrete on the exposed surface. The surface layer of concrete affects the transport characteristics of the material and can enhance the chloride ingress. Also, although a more complex study must be performed, based on the available data, it was found that carbonation influences the chloride front and decreases significantly the chloride concentration. Even though having a limited number of replicas, a statistical analysis at a confidence level of 95% was realized in order to investigate the possible influence of crack width on chloride ingress, but no statistically significant influence was found within the range considered. Chapter 4 presents a numerical model that can realistically simulate chloride ingress in uncracked and cracked concrete. Initially, a 2D model was developed to simulate chloride transport on mortar samples with artificial cracks (notches), with different widths and depths, using the Abaqus/Standard software based on the finite element method (FEM) using the mass diffusion tool. The algorithm for solving mass diffusion equations in Abaqus is described. The numerical results obtained were compared to the experimental ones and a good agreement between them was found, showing the validity of the proposed model. Also, from these simulations it was found that within the parameter range considered the chloride transport property of concrete is not influenced by the considered crack widths, but only by the crack depths. Furthermore, it was also concluded that chloride penetration depth has a reduced sensitivity to changes in the diffusion coefficient, when simulating chloride diffusion. A 3D model was then used to simulate chloride ingress in the uncracked and cracked samples used in the experimental part previously described. It must be mentioned that as in the case of the 2D model presented above, the model parameters such as the diffusion coefficient D, the initial chloride concentration c and the applied chloride concentration C, were determined experimentally and were used as simulation parameters. Chloride ingress was simulated in uncracked concrete (reference samples type S) and afterwards it was simulated on cracked concrete (Sample 4-of type SC with cracks and without rebars). When comparing the numerical results with the experimental ones, good agreement was found. Also, this chapter presents a 3D model of each sample used in the experimental part, with regards to the individual geometrical characteristics of each sample, including the crack pattern and rebar position. In chapter 5, final conclusions, personal contributions and some suggestions of further research are presented

from #MedicinebyAlexandrosSfakianakis via xlomafota13 on Inoreader http://ift.tt/1Np5vk2
via IFTTT